The Human F-Box DNA Helicase FBH1 Faces Saccharomyces cerevisiae Srs2 and Postreplication Repair Pathway Roles

The Saccharomyces cerevisiae Srs2 UvrD DNA helicase controlsgenome integrity by preventing unscheduled recombination events.While Srs2 orthologues have been identified in prokaryotic andlower eukaryotic organisms, human orthologues of Srs2 have notbeen described so far. We found that the human F-box DNA helicasehFBH1 suppresses specific recombination defects of S. cerevisiaesrs2 mutants, consistent with the finding that the helicasedomain of hFBH1 is highly conserved with that of Srs2. Surprisingly,hFBH1 in the absence of SRS2 also suppresses the DNA damagesensitivity caused by inactivation of postreplication repair-dependentfunctions leading to PCNA ubiquitylation. The F-box domain ofhFBH1, which is not present in Srs2, is crucial for hFBH1 functionsin substituting for Srs2 and postreplication repair factors.Furthermore, our findings indicate that an intact F-box domain,acting as an SCF ubiquitin ligase, is required for the DNA damage-induceddegradation of hFBH1 itself. Overall, our findings suggest thatthe hFBH1 helicase is a functional human orthologue of buddingyeast Srs2 that also possesses self-regulation properties necessaryto execute its recombination functions.

INTRODUCTION

Top

INTRODUCTION

DNA lesions occur frequently in living cells as a result ofspontaneous events or external insults. Growing evidence suggeststhat the selection of the appropriate DNA repair pathway todeal with broken DNA molecules is crucial for preventing genomeinstability.

The Saccharomyces cerevisiae Srs2 protein is a 3'-5' DNA helicase(40) structurally and functionally related to bacterial UvrD(2, 51). It is thought that Srs2 plays a key role in influencingthe choice between homologous recombination (HR) and postreplicationrepair (PRR) pathways, both of which are required to counteractthe accumulation of gaps during DNA replication (5, 45). A bodyof evidence suggests that Srs2 inhibits HR at an early step(1, 7, 11, 20, 29, 39, 44), acting as a DNA translocase thatdisassembles the Rad51 nucleofilament (26, 52). Accordingly,srs2 ${Delta}$ mutants show hyperactivation of spontaneous recombinationevents (41), and unrestrained HR is the source of cell deathin srs2 mutants when other factors operating at later stagesin recombination are also inactivated. This is the case forSgs1 (14) and Rad54 (23) helicases, which are involved in theresolution of mature recombination intermediates and in promotingD-loop formation and/or stabilization (47, 55), respectively.

Current models indicate that Srs2 inhibits HR and channels DNAlesions towards the PRR pathway. This model is mainly supportedby the observation that the DNA damage sensitivity of PRR mutantscan be rescued by SRS2 inactivation in the presence of a functionalHR pathway (5, 45). The PRR pathway seems to be required totolerate rather than immediately repair the DNA damage, usingboth specialized translesion synthesis DNA polymerases and aless-characterized error-free repair branch that is thoughtto involve a recombination-dependent replication mechanism,such as template switching (5). The PRR pathway is promotedby two diverse protein complexes that contain ubiquitin-conjugating(E2) and ubiquitin ligase (E3) enzymes and that modify PCNAon lysine (K) 164. In particular, the Rad6 (E2)-Rad18 (E3) complexmono-ubiquitylates PCNA on K164, while Rad5 (E3), along withthe heterodimeric Mms2-Ubc13 (E2 variant) enzyme, subsequentlyattaches polyubiquitin chains via K63 to the mono-ubiquitylatedK164 residue (18). Polyubiquitin chains linked by K63 isopeptidebonds, in contrast to those assembled in K48 conformation, donot mark the modified proteins for degradation but, rather,signal for other cellular transactions (37). The K164 residueof PCNA can also be sumoylated by a Ubc9/Siz1-dependent pathway(18). Recent evidence obtained with budding yeast indicatesthat the modification status of PCNA is crucial in determiningthe PRR subpathway that will be engaged in front of a lesion.In particular, the Srs2 antirecombinogenic function is enhancedby its physical interaction with sumoylated PCNA (34, 36), sumoylationand mono-ubiquitylation of PCNA contribute to spontaneous mutagenesismediated by translesion DNA polymerases (46), and the polyubiquitylationof PCNA might promote the error-free PRR subpathway (18). Inspite of the suggestion that sumoylated PCNA, by recruitingSrs2 at replication forks, would play a key role in preventingHR during S phase (34, 36), recent studies failed to detectthe accumulation of recombination intermediates at damaged replicationforks in srs2 or even PCNA mutants, in which sumoylation and/orubiquitylation is abrogated (4, 29). On the other hand, recombinationintermediates broadly accumulate at damaged forks in sgs1 orubc9/mms21 mutants (4, 29), suggesting that the control of HRduring replication is far more complicated than the sole regulationof Srs2. Moreover, a number of srs2 recombination phenotypescannot easily be explained by considering Srs2 only as a proteinthat counteracts Rad51-mediated strand invasion. Rather, manygenetic findings suggest that Srs2 operates at diverse levelsin recombination, downstream of the strand invasion step (39),and possibly in promoting double-strand-break (DSB) repair byHR (3, 19, 20, 35).

Despite its key role in maintaining genome integrity in yeast,Srs2 has no human orthologues identified so far. However, theSrs2 DNA helicase has been conserved in other fungi, such asSchizosaccharomyces pombe and Neurospora crassa, and the geneticcharacterization of null mutants in both organisms showed thatthe Srs2 orthologues are fundamental players in HR (10, 30,31, 48, 53). Genetic analysis carried out with fission yeastsuggests that Srs2 has overlapping functions in processing recombinationintermediates with both Rqh1 and Fbh1 DNA helicases (32, 33).Fbh1 orthologues have been found in mice, chickens, and humansbut not in budding yeast (21, 24). In DT40 cells, Fbh1 and BLMproteins cooperate in constraining the extent of sister chromatidexchanges (SCEs) in HR (24). Unusually for DNA unwinding enzymes,Fbh1 DNA helicases possess an F-box motif (22). Studies performedon human cell extracts indicated that human FBH1 (hFBH1), throughthe F-box domain, is part of a Skp-Cullin-F-box (SCF) ubiquitinligase complex; while it has been shown that the E3 SCF complexcontaining hFBH1 promotes the assembly of polyubiquitin chains,the ubiquitin targets of hFBH1 are still unknown (22).

The aim of our study was to identify a human orthologue forbudding yeast Srs2. Here we show that the F-box DNA helicasehFBH1 suppresses specific recombination defects of S. cerevisiaesrs2 mutants. Furthermore, when hFBH1 substitutes for Srs2,the PRR functions required to induce the K164 ubiquitylationof PCNA are dispensable for cell survival upon DNA damage treatment.The F-box domain is essential for hFBH1 functions in substitutingfor Srs2 and PRR roles, and it controls the DNA damage-inducedturnover of hFBH1 by cooperating with yeast SCF components.Overall, our data indicate that hFBH1 represents a bona fidefunctional orthologue of budding yeast Srs2 that, in addition,during its evolution might have acquired self-regulatory propertiesnecessary to modulate its DNA recombination functions.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Plasmids, yeast strains, and procedures. All of the yeast strains used in this study are isogenic tothe W303 background. Deletion or Myc-tagged strains were obtainedby a one-step method. Multiple mutant strains were derived bymeiotic segregants. The srs2::hFBH1 strain was constructed asfollows. The HIS3-MX6 cassette was inserted at the C terminusof the hFBH1 gene carried by the pIC1 plasmid, a derivativeof pAS2-1 (21), thus creating the pIC1-HIS3 plasmid. The plasmidregion containing both hFBH1 and HIS3 was amplified by PCR,using primers with 5' and 3' tails homologous to the genomicsequences flanking the genomic SRS2 locus. The resulting PCRproduct was used to transform a wild-type (wt) W303 yeast strain.Details about the oligonucleotide sequences are available uponrequest. hfbh1- ${Delta}$ N and hfbh1-hd strains were constructed usingthe same strategy by amplifying a PCR fragment lacking the F-boxmotif or a PCR fragment containing a K520A mutation obtainedby site-direct mutagenesis of plasmid pIC1-HIS3. The F-SRS2strain was constructed by inserting the hFBH1 F-box domain intothe NotI site of the pG35 plasmid (28) to generate the pSAPO1plasmid. PstI-linearized pSAPO1 plasmid was used to integratethe F-SRS2 construct at the SRS2 locus.

Standard genetic analyses were performed according to publishedprocedures (42). An intrachromosomal recombination assay wasperformed using yeast strains carrying a heteroallelic duplicationof LEU2, with URA3 between the LEU2 genes, as previously described(25). Spot assays were performed by evaluating cellular growthon synthetic complete medium containing adenine at a final concentrationof 0.7 g/liter, with or without methyl methanesulfonate (MMS),hydroxyurea (HU), and 4-nitroquinolone-1-oxide (4-NQO; Sigma).All spot tests were repeated at least three times in independentexperiments, using different yeast transformants or segregantsto ensure the reproducibility of the results. Dose-responsekilling curves for zeocin (Invitrogen) were determined by platingserial dilutions of exponentially growing cells treated for1 hour with or without the mutagen; to measure UV light sensitivity,a known number of cells were plated from exponentially growingcultures and exposed to different UV doses. CFU were evaluatedafter 3 to 5 days at 28°C. All survival curves were repeatedthree times in independent experiments. Sodium dodecyl sulfate-polyacrylamidegel electrophoresis and Western blot procedures were done aspreviously described (28), using anti-Myc 9E11 (Bio Optica,United Kingdom) and ${alpha}$ -tubulin (Oxford Biotechnology, United Kingdom)antibodies. Nocodazole and cycloheximide (Sigma) were used atfinal concentrations of 10 µg/ml and 50 µg/ml, respectively.The protein quantification analysis shown in Fig. 5 was carriedout using Image J (NIH) software. The relative amounts of thehFBH1 protein were normalized by using tubulin as a loadingcontrol, as calculated in independent experiments.

View larger version (20K):
[in this window]
[in a new window]

FIG. 5. hFBH1 protein turnover is stimulated by DNA damage and depends on a functional F-box domain and yeast SCF complex. (A) Different cell cultures of an hFBH1 Myc-tagged yeast strain were treated for 3 h at the indicated MMS concentrations. Cycloheximide (chx) was then added to each culture to prevent further protein synthesis. At the indicated times, crude protein extracts were prepared and analyzed by Western blotting using anti-Myc or anti-tubulin antibody as a loading control. Quantification analysis of hFBH1 protein levels is also shown. (B) Log-phase cells of the indicated yeast strains were presynchronized by nocodazole treatment, released at 25°C into 0.02% MMS for 2 h, and then incubated at 37°C for 1.5 h. Cycloheximide was subsequently added, and protein samples were analyzed at the indicated time points as described for panel A. hFBH1 and F-box-truncated mutant proteins are indicated by asterisks. Quantification analysis of hFBH1 protein levels is also shown.

RESULTS

Top

RESULTS

The S. cerevisiae Srs2 protein contains seven segments of aminoacids that constitute the helicase motifs and are well conservedin all DNA helicases of the UvrD family, including Escherichiacoli UvrD, S. pombe Srs2, and N. crassa Srs2 (2, 48, 53). Tosearch for putative human orthologues of S. cerevisiae Srs2,the helicase motifs were compared by a BLAST search of the Homosapiens genome. The F-box DNA helicase hFBH1 (21) scored thebest positive result in our analysis. As shown in Fig. 1A, sequencealignment of the helicase motifs of hFBH1 with those of S. cerevisiaeSrs2, S. pombe Srs2, and bacterial UvrD reveals significantconservation, which is limited to the helicase domains. In particular,the helicase domains of S. cerevisiae Srs2 and hFBH1 possess19.2% identity and 34.4% similarity at the amino acid level.S. cerevisiae Srs2 is the only UvrD member with a long C-terminaltail (Fig. 1B), which has been implicated in protein regulation(8, 26, 36, 40). Conversely, the Fbh1 proteins in S. pombe,mice, and humans are characterized by the presence of an F-boxdomain, located in the N-terminal part, that is absent in Srs2and in bacterial UvrD helicases (Fig. 1B).

View larger version (21K):
[in this window]
[in a new window]

FIG. 1. The UvrD helicase family. (A) Amino acid sequence alignments, performed with CLUSTALW, among the seven helicase motifs found in E. coli (ec) UvrD, S. cerevisiae (sc) Srs2, S. pombe (sp) Srs2, S. pombe Fbh1, mouse (m) Fbox18, and hFBH1. Identical and conserved amino acids are indicated within gray and white boxes, respectively. (B) Schematic representation of UvrD helicases and conserved motifs.

To address whether hFBH1 might represent a functional orthologueof S. cerevisiae Srs2, we tested the ability of hFBH1 to suppressspecific phenotypes of srs2 ${Delta}$ mutants. For this analysis, we replacedthe SRS2 open reading frame at its chromosomal location withthe human cDNA encoding the hFBH1 protein. In the resultingsrs2::hFBH1 strain, the yeast SRS2 promoter, which is responsiblefor the cell cycle and the DNA damage-induced regulation ofthe original transcript (16), controls the expression of thehFBH1 cDNA insert. By using Western blot analysis, we foundthat the hFBH1 protein is expressed in the yeast strain, althoughits level is lower than the level of the endogenous Srs2 protein(see Fig. 3C). srs2 ${Delta}$ mutants are characterized by hyperactivationof spontaneous recombination (25, 41), and such a deletion islethal in the absence of RAD54 (23) or SGS1 (14), which areboth required for the maturation of recombination intermediates.The cell lethality of srs2 rad54 and srs2 sgs1 double mutantsis suppressed by deleting RAD51 (14, 23, 44), consistent withthe idea that Srs2 dismantles Rad51 nucleofilaments (26, 52).Using a genetic system that allows the analysis of gene conversionand recombination-mediated deletion events occurring betweendirected repeats, we confirmed previous observations (25) indicatingthat in srs2 ${Delta}$ mutants the rate of gene conversion is enhancedfour to five times compared to that of wt cells. We found thatthe gene conversion rate was reduced in the srs2::hFBH1 straincompared to that of the srs2 ${Delta}$ strain (Fig. 2A), indicating thathFBH1 is able to counteract recombination in the absence ofSrs2. We also found, by tetrad analysis, that while srs2 ${Delta}$ mutantsare synthetically lethal in combination with rad54 ${Delta}$ deletion,srs2::hFBH1 rad54 ${Delta}$ double mutants are viable (Fig. 2B). We thentested the ability of hFBH1 to prevent cell lethality in srs2 ${Delta}$ sgs1 ${Delta}$ double mutants and found that the presence of a wt copyof SRS2, carried by a URA3 plasmid and counterselectable bytreatment with 5-fluoroorotic acid, is much less critical forcell survival in srs2::hFBH1 sgs1 ${Delta}$ mutants than in srs2 ${Delta}$ sgs1 ${Delta}$ mutants (Fig. 2C). srs2 ${Delta}$ , but not sgs1 ${Delta}$ , mutants require RAD27for cell viability (9, 23). RAD27 encodes a Flap endonucleaseimplicated in processing of Okazaki fragments during DNA replication,and rad27 mutants also exhibit synthetic lethality in combinationwith HR mutants (49). These genetic observations might implythat the DNA lesions generated by the absence of Rad27 are repairedmainly through HR and are not in contrast with the possibilitythat Srs2 favors recombinational repair in this context. rad27 ${Delta}$ cells were crossed with srs2 ${Delta}$ or srs2::hFBH1 cells, and tetradanalysis revealed that while srs2 ${Delta}$ rad27 ${Delta}$ double mutants are syntheticallylethal, srs2::hFBH1 rad27 ${Delta}$ cells are viable (Fig. 2D). We thereforeconcluded that hFBH1 can also substitute for Srs2 in processingDNA lesions that spontaneously arise in rad27 ${Delta}$ mutants. Takentogether, the data in Fig. 2 indicate that hFBH1 suppressesthe spontaneous recombination defects occurring in the absenceof a functional Srs2 protein.

View larger version (31K):
[in this window]
[in a new window]

FIG. 3. hFBH1 suppresses the hypersensitivity of srs2 mutants to mutagens. (A) The indicated strains were grown at an equal cellular concentration, sequentially diluted 1:6, and spotted onto plates containing MMS, HU, or 4-NQO at the indicated concentrations. Cellular growth was evaluated after incubation at 28°C for 3 to 5 days. (B) An appropriate number of log-phase cells of wt, srs2 ${Delta}$ , and srs2::hFBH1 yeast strains were plated and exposed to different doses of UV light. Cell survival was then evaluated compared to that of an untreated control. Log-phase cultures of the same strains were incubated with different doses of zeocin for 1 h, and cell survival was then calculated as the plating efficiency with respect to untreated cells. (C) Schematic view of hFBH1, Srs2, and mutants used in this study. Crude protein extracts were prepared from the indicated Myc-tagged strains grown under unperturbed conditions (–) or in the presence of 0.02% MMS for 3 h (+) and were analyzed by Western blotting using antibodies against Myc or ${alpha}$ -tubulin as a loading control.

View larger version (21K):
[in this window]
[in a new window]

FIG. 2. hFBH1 suppresses srs2 recombination defects arising from spontaneous DNA lesions. (A) Gene conversion rates were determined for wt, srs2 ${Delta}$ , and srs2::hFBH1 strains. (B) Tetrads obtained from sporulation of diploids heterozygous for rad54 ${Delta}$ and srs2 ${Delta}$ (left) or rad54 ${Delta}$ and srs2::hFBH1 (right). (C) The indicated strains, containing SRS2 cloned into a URA3⁺ vector, were grown at an equal cellular concentration and sequentially diluted 1:6 before being spotted onto plates without treatment (UNT) or with 5-fluoroorotic acid (5-FOA) to counterselect the plasmid. Cellular growth was evaluated after incubation at 28°C for 3 to 5 days. (D) Tetrads obtained from sporulation of diploids heterozygous for rad27 ${Delta}$ and srs2 ${Delta}$ (left) or rad27 ${Delta}$ and srs2::hFBH1 (right).

srs2 mutants exhibit hypersensitivity when exposed to a varietyof DNA-damaging agents that cause the accumulation of differentDNA lesions, including DSBs (45). We used a colony formationassay to test whether hFBH1 was able to rescue the hypersensitivityof srs2 ${Delta}$ cells to different mutagens. As shown in Fig. 3A, wefound that hFBH1 suppresses the hypersensitivity of srs2 ${Delta}$ mutantsto MMS, HU, and 4-NQO. Similarly, hFBH1 suppresses srs2 ${Delta}$ hypersensitivityto UV light and zeocin, as assessed by treating exponentiallygrowing cells with increasing doses of mutagens (Fig. 3B).

We then dissected the contributions of the helicase and F-boxdomains to hFBH1 function by creating two specific mutants (Fig.3C) in which the F-box motif was deleted (hfbh1- ${Delta}$ N) or in whichthe conserved K within the ATP binding domain, essential forthe enzymatic activities of the yeast Srs2 protein (25), waschanged to A (hfbh1-hd). Both hfbh1 alleles were replaced atthe SRS2 locus, and using Western blot analysis, we found thatthe mutated proteins were expressed at least at the same levelas wt hFBH1, both under untreated conditions and in responseto MMS (Fig. 3C). srs2::hfbh1- ${Delta}$ N and srs2::hfbh1-hd strains showedincreased sensitivity to drug treatments, resembling that ofsrs2 ${Delta}$ mutants, compared to the srs2::hFBH1 strain (Fig. 3A).Hence, both the helicase and the ubiquitin ligase functionsof hFBH1 are essential for suppressing the DNA damage sensitivityof srs2 ${Delta}$ mutants. Moreover, the DNA damage sensitivities of rad51 ${Delta}$ ,srs2 ${Delta}$ rad51 ${Delta}$ , and srs2::hFBH1 rad51 ${Delta}$ strains were comparable, indicatingthat hFBH1 suppresses srs2 ${Delta}$ DNA damage sensitivity only in thepresence of a functional HR pathway (Fig. 3A).

It is thought that Srs2 channels DNA lesions into the PRR pathwayby preventing HR (5, 45). According to this model, the elevatedDNA damage sensitivities of PRR-deficient mutants can be rescuedpartially by deleting SRS2 in the presence of a functional HRpathway (2, 6, 13, 41, 43, 50). We therefore tested whetherhFBH1 in the absence of Srs2 could restore the MMS sensitivityof PRR-deficient mutants. As previously shown, we found thatthe deletion of SRS2 partially rescued the MMS sensitivitiesof rad5 ${Delta}$ , rad18 ${Delta}$ , rad6 ${Delta}$ , and ubc13 ${Delta}$ mutants (Fig. 4A); surprisingly,we found that rad5 ${Delta}$ srs2::hFBH1, rad18 ${Delta}$ srs2::hFBH1, and ubc13 ${Delta}$ srs2::hFBH1 mutants were more resistant to MMS treatment thanrad5 ${Delta}$ srs2 ${Delta}$ , rad18 ${Delta}$ srs2 ${Delta}$ , and ubc13 ${Delta}$ srs2 ${Delta}$ mutants, while rad6 ${Delta}$ srs2::hFBH1cells were more sensitive to MMS than rad6 ${Delta}$ srs2 ${Delta}$ mutants (Fig.4A). In response to DNA damage, the PRR factors deleted in theabove mutants induce the mono- and polyubiquitylation of PCNAat the conserved K164 residue; srs2 ${Delta}$ mutants rescue the DNA damagesensitivity of ubiquitylation-defective pcnaK164R mutants (18).We found that hFBH1 also rescues the MMS sensitivity of pcnaK164Rmutants in the absence of Srs2 and, similar to what was foundfor rad5 ${Delta}$ , rad18 ${Delta}$ , and ubc13 ${Delta}$ mutants, that this rescue is betterthan that by deletion of SRS2 itself, although not to the wtlevel (Fig. 4A). Thus, ubiquitylation of PCNA at K164 is alsodispensable for MMS survival when hFBH1 substitutes for Srs2.

View larger version (40K):
[in this window]
[in a new window]

FIG. 4. hFBH1 and the hybrid F-Srs2 protein suppress the DNA damage sensitivity of PRR mutants. (A and B) The indicated strains were grown at equal cellular concentrations, sequentially diluted 1:6, and spotted onto plates containing MMS at the indicated concentrations. Cellular growth was evaluated after incubation at 28°C for 3 to 5 days.

Altogether, these findings indicate that hFBH1 compensates notonly for the absence of Srs2 in promoting cell survival uponMMS treatment but also for PRR functions related to PCNA ubiquitylation,although not for those of Rad6. To further characterize thisaspect, we asked if the F-box domain of hFBH1 was required forthe PRR suppression phenotype. Indeed, we found that fbh1- ${Delta}$ Ndid not rescue the MMS sensitivity of rad5 ${Delta}$ mutants (Fig. 4B).Notably, we found that hfbh1-hd mutants behaved similarly tohfbh1- ${Delta}$ N mutants (Fig. 4B), suggesting that a functional hFBH1helicase domain is also required for rad5 ${Delta}$ suppression. Sincewe cannot rule out that deletion of the F-box domain partiallyaffects the helicase activity of hFBH1 (22), we tested the contributionof the F-box domain to the suppression of MMS sensitivity ofPRR-deficient mutants in a different way. We asked if the additionof the hFBH1 F-box domain to the yeast Srs2 protein could besufficient to relieve the need for PRR factors. To test thishypothesis, we created a hybrid SRS2 gene carrying the codingregion for the F-box domain of hFBH1 (Fig. 3C) and used thischimera to replace endogenous SRS2 in PRR-deficient backgrounds.We found that the F-box-Srs2 protein (F-Srs2) restored the MMSsensitivity of rad5 ${Delta}$ or rad18 ${Delta}$ cells to a similar extent to thatby hFBH1 (Fig. 4B). Hence, the hybrid F-box Srs2 protein recapitulateshFBH1 properties in suppressing PRR mutants.

We noted that the MMS sensitivity of rad5 ${Delta}$ srs2::hFBH1rad51 ${Delta}$ cellswas identical to that of rad5 ${Delta}$ srs2 ${Delta}$ rad51 ${Delta}$ or rad5 ${Delta}$ rad51 ${Delta}$ cells,indicating that the suppression of the DNA damage sensitivityof rad5 ${Delta}$ mutants by hFBH1 requires a functional RAD51 gene (Fig.4B). Conversely, neither the inactivation of RAD30 nor thatof REV3, resulting in ablation of the translesion DNA polymerase ${eta}$ and DNA polymerase ${zeta}$ , respectively, affected the capabilityof hFBH1 to suppress the MMS sensitivity of rad5 mutants inthe absence of SRS2 (data not shown).

Altogether, the data presented in Fig. 4 indicate that the replacementof Srs2 with hFBH1 suppresses PRR mutant repair defects in thepresence of an intact HR pathway and that the F-box domain,in association with functional UvrD helicase activity, playsan essential role in this suppression.

In the course of our studies, we noticed that the deletion ofthe F-box domain in hfbh1- ${Delta}$ N mutants led to an increase of thehFBH1 protein level, while its addition to endogenous Srs2 inF-SRS2 strains correlated with a reduced Srs2 protein level(Fig. 3C). In order to better characterize this aspect, we measuredhFBH1 protein stability in yeast strains, using the proteinsynthesis inhibitor cycloheximide.

Different cultures of a yeast strain carrying a Myc-tagged versionof hFBH1 were treated with increasing doses of MMS and thenexposed to cycloheximide to prevent further protein synthesis.As shown in Fig. 5A, the hFBH1 protein level abruptly decreasedin time upon protein synthesis inhibition, and in particular,the hFBH1 turnover rate increased by raising the dose of MMS,as also judged by quantification analysis. Hence, although ouranalysis suggests that hFBH1 is subjected to rapid degradationeven under unperturbed conditions (data not shown), the DNAdamage treatment further stimulated its turnover. In contrast,the turnover of a truncated version of hFBH1 lacking the F-boxmotif was dramatically impaired in MMS-treated cells (Fig. 5B).Given that hFBH1 interacts with Skp1 through the F-box domainto form an SCF complex in vitro (22), we asked if yeast SCFfactors could be required for hFBH1 protein turnover under ourexperimental conditions. As shown in Fig. 5B, we found thathFBH1 degradation is indeed prevented by inactivating the Cdc53or Skp1 SCF component by using temperature-sensitive alleles.

Overall, these findings indicate that hFBH1 is subject to rapidturnover, which is further stimulated by DNA damage treatment,and that this turnover depends on a functional F-box domainand yeast SCF-dependent degradation machinery.

DISCUSSION

Top

DISCUSSION

Several human inherited genetic diseases, characterized by genomicinstability and cancer predisposition, are due to mutationsin genes encoding DNA helicases, a class of enzymes that controlfundamental aspects of DNA metabolism, including DNA replication,repair, recombination, and transcription. In budding yeast,the Srs2 DNA helicase plays a crucial role in the maintenanceof genome stability by regulating DNA recombination. Srs2 hasbeen conserved structurally and functionally in bacteria andin other fungi, such as S. pombe and N. crassa (2, 10, 30, 48,51, 53), yet a human orthologue of Srs2 has never been described,thus leading to the idea that its functions might be sustainedby other structurally unrelated DNA helicases in mammalian cells(17).

hFBH1 is highly similar to Srs2 within the helicase domain, but it also carries an F-box motif that self-regulates its turnover. With the aim of searching for putative human orthologues ofS. cerevisiae Srs2, we compared by BLAST searching just itsconserved UvrD helicase domain with the H. sapiens genome andidentified the F-box DNA helicase hFBH1 as the protein withthe best score. Indeed, within the helicase domain, hFBH1 sharesa high degree of similarity with all members of the UvrD proteinfamily, including S. pombe Srs2 and E. coli UvrD. Beyond thehelicase domain, hFBH1 shares poor homology with UvrD proteins,and it exclusively contains an F-box domain. Bona fide orthologuesof hFBH1 have been identified in S. pombe, mouse, and chickencells (21, 24). Considering the similarities and differencesbetween Srs2 and Fbh1 protein subgroups, they likely constitutea single protein family, resembling the case of RecQ helicases,where the different members share homology mainly within theirhelicase domains (17). Notably, for S. pombe, members of thetwo subgroups have been described, and genetic data suggestthat fbh1 ${Delta}$ srs2 ${Delta}$ double mutants are lethal due to unrestrainedHR (32, 33). In particular, it has been suggested that S. pombeFbh1 has a role in disassembling Rad51 nucleofilaments, eitherthose formed in the absence of HR mediator proteins (33) orthose created after strand invasion (32). Similarly, in DT40chicken cells, Fbh1, by cooperating with the BLM RecQ helicase,acts in limiting SCEs (24). Thus, in different organisms, Fbh1functions rely on recombination, and our findings indicate thathFBH1 can functionally substitute for budding yeast Srs2.

As mentioned above, hFBH1 has an F-box domain, a feature notshared with S. cerevisiae Srs2. The F-box proteins are the interchangeablerecognition subunits of ubiquitin ligase SCF complexes thatcontrol protein degradation in different cellular processes(54). In vitro studies indeed indicate that hFBH1 is part ofan SCF complex that promotes the formation of ubiquitin chains,although the targets modified in vivo are not known (22). Consistentwith findings obtained by studying the S. pombe Fbh1 counterpart(32, 33), our data indicate that the F-box domain is necessaryfor hFBH1 function(s) in substituting for Srs2, and in addition,they raise the possibility that one target of the F-box domainis hFBH1 itself. Indeed, based on data from yeast, hFBH1 isan unstable protein whose turnover is further stimulated byDNA damage and depends on a functional F-box domain and SCFcomponents.

hFBH1 fulfils S. cerevisiae Srs2 functions in HR. Despite the presence of the F-box motif, does hFBH1 accomplishthe specific Srs2 recombination functions? In fact, our dataindicate that hFBH1 suppresses many srs2-specific recombinationdefects, thus indicating that hFBH1 is able to substitute forSrs2 in processing recombination intermediates in differentcontexts. In particular, it has been suggested that certainsrs2 phenotypes, including synthetic lethality with rad54 orsgs1 mutants and UV hypersensitivity, can be ascribed to theaccumulation of toxic Rad51 nucleofilaments (1, 14, 23), a hypothesissustained by the finding that Srs2 displaces Rad51 nucleofilamentsin vitro (26, 52). It is not known whether Fbh1 proteins havesuch DNA translocase activity, but since they are members ofthe UvrD family, it would not be surprising. In addition, Srs2is also required for DSB repair by HR (3, 19, 20, 35), a rolethat might also explain the hypersensitivity of srs2 mutantsto DNA-damaging agents that cause DSBs or their synthetic lethalitywith rad27 mutants. In these contexts, Srs2 might act in favoringrecombinational repair. It was suggested that the prorecombinationalrole of Srs2 still reflects its ability to dislodge aberrantor unscheduled Rad51 nucleofilaments that might prevent recombinationalrepair (1, 33). Otherwise, Srs2 and Fbh1 might also counteractRad51 nucleofilaments after strand invasion, thus constrainingSCE events to the level necessary to repair interrupted DNAmolecules. Notably, a role for S. cerevisiae Srs2 and even forS. pombe Fbh1 downstream of the Rad51 strand invasion step issupported by genetic findings (20, 32, 39). Finally, since bothSrs2 and Fbh1 unwind the DNA duplex in vitro as proper DNA helicases(21, 40), they might have a more direct role in vivo in promotingrecombination by extending the D loop once strand invasion hasoccurred (35).

hFBH1 faces error-free PRR functions by the aim of its F box. What is more surprising in our findings is that hFBH1, in theabsence of SRS2, suppresses the DNA damage sensitivity of PRRmutants, while in contrast, Srs2 causes cell death of PRR mutants(5, 45). As a consequence, hFBH1 and Srs2 have apparently oppositeeffects on the recombination outcome in this context: whileSrs2 is thought to kill PRR mutants by preventing recombination,hFBH1 might rescue PRR mutants from cell lethality by favoringrecombination. The requirement of HR for this suppression mightsupport this possibility. Notably, the suppression of PRR mutantcell lethality by hFBH1 cannot be ascribed to the observationthat hFBH1 might be present in limiting amounts with respectto Srs2. In fact, hFBH1 suppresses the MMS sensitivity of PRRmutants even more efficiently than that of srs2 ${Delta}$ mutants. Ourdata rather suggest that this suppression entails an activeUvrD helicase whose turnover depends on its F-box motif, resultingin an enzyme that efficiently favors recombinational repair.

According to the current model, Srs2 exerts antirecombinationfunctions when it is recruited by sumoylated PCNA (34, 36).The association of Srs2 with sumoylated PCNA depends on itsC-terminal tail, and its removal induces a PRR suppression phenotyperesembling the one caused by preventing PCNA sumoylation (36).Although the C-terminal domain of Srs2 has not been conservedin hFBH1, our data indicate that the hybrid F-Srs2 protein thatretains the SUMO-PCNA binding domain (Fig. 3C) is still ableto suppress the MMS sensitivity of PRR mutants. Furthermore,both F-Srs2 and hFBH1 suppress PRR mutants better than mutationsthat prevent the sumoylation of PCNA, whose phenotypes in thisregard resemble those of srs2 ${Delta}$ mutants (36). Thus, the presenceof the F-box domain rather than the absence of the C-terminaltail seems to be crucial for the suppression of PRR mutantsby hFBH1 or F-Srs2.

Although other possibilities could be envisaged, we favor oneinterpretation of our data that might suggest a novel functionalrelationship between Srs2 and PRR factors in S. cerevisiae,as follows. In response to DNA damage, Srs2 might act mainlyin favoring recombinational repair and PRR factors might enhancethis Srs2 prorecombinational activity. In PRR mutants, Srs2becomes toxic because it is unable to stimulate proper recombinationevents, and it simultaneously prevents access to DNA lesionsof other factors that promote HR in its absence. In this scenario,hFBH1, by the aim of its F-box, fulfils Srs2 functions in HRand the PRR regulatory role in Srs2 activity at the same time.

But how could PRR genes influence Srs2 duties, and what is themechanism by which the F-box hFBH1 helicase could substitutefor PRR functions? Studies with human cells showed that hFBH1is a component of the SCF ubiquitin ligase complex (22), andSCF complexes have been implicated in K48 ubiquitylation andin 26S proteasome-dependent degradation of their targets (54).Therefore, one likely possibility is that hFBH1 disruption occursthrough an autoubiquitylation mechanism (Fig. 6). Since theF-box domain self-regulates the turnover of hFBH1 and is requiredfor the suppression of PRR mutants, one interesting hypothesisis that the control exerted by the PRR pathway on Srs2 relieson a similar mechanism. Rad6/Rad18 and Rad5/Ubc13 cooperateto assemble ubiquitin via K63 chains, which is not a typicalsignal for protein degradation (37); thus, it is unlikely thatthe PRR pathway would directly promote Srs2 disruption. However,in response to DNA damage, the PRR pathway promotes PCNA ubiquitylation(34, 36). Given that hFBH1 also rescues ubiquitin-defectivepcnaK164R mutants, an intriguing possibility is that one ofthe essential functions of PRR factors, by inducing PCNA modification,is in turn to influence the stability of Srs2 (Fig. 6). Thishypothesis might not be entirely surprising, since it was recentlyshown that the replication factor Cdt1 is ubiquitylated anddegraded, also in response to DNA damage, once it is recruitedby PCNA to the DNA template (15). In light of this, we speculatethat Srs2 degradation in its natural context is promoted byan SCF complex and that an unidentified endogenous F-box proteinin yeast could perform the same F-box-dependent function ofhFBH1 (Fig. 6). Therefore, the PRR pathway, by inducing themodification of PCNA, might at the same time influence the capabilityof an SCF complex to disrupt Srs2 and possibly other factorsbound to DNA, maybe by stimulating the recruitment of a putativeF-box protein. In this view, hFBH1 does not require certainPRR functions to survive DNA damage because it is able to self-regulateits turnover. Similarly, the proper turnover of endogenous Srs2in PRR mutants can be reestablished by adding the sole F-boxdomain of hFBH1 to the helicase.

View larger version (19K):
[in this window]
[in a new window]

FIG. 6. Model for SCF-mediated regulation of hFBH1 and Srs2 turnover in response to DNA damage. The F-box domain of hFBH1 forms an SCF ubiquitin ligase that controls hFBH1 DNA damage-induced turnover. In budding yeast under unperturbed conditions, Srs2 is recruited by sumoylated PCNA, and in response to DNA damage, the conserved Rad6/Rad18/Rad5/Ubc13 PRR pathway triggers K63 ubiquitylation of PCNA. PCNA ubiquitylation stimulates the recruitment of a putative SCF complex that, as in the case of the hFBH1 F-box domain, enhances Srs2 turnover. In addition to its role in monoubiquitylating PCNA, Rad6, by promoting the modification of another unknown target(s), aids hFBH1 and Srs2 recombination functions. Increased hFBH1 and Srs2 turnover might lead to the stimulation of recombinational repair of DNA lesions.

Our data also suggest that hFBH1 suppresses the MMS sensitivityof PRR mutants, with the exception of rad6 mutants. AlthoughRad6 has E2 activity (27), SCF ubiquitin ligases work preferentiallywith the Cdc34 E2 enzyme (54); in fact, we found that in rad6mutants, hFBH1 degradation is not prevented (data not shown).Thus, to explain our suppression data, we favor the hypothesisthat Rad6 might control the modification of another target(s)necessary for hFBH1 function in favoring recombinational repair.Rad6 cooperates with E3 ligases other than Rad18, and by targeting,for example, histones (27), it might aid hFBH1 in processingrecombination intermediates.

Concluding remarks. Genetic analysis performed with fission yeast suggests thatthe two UvrD helicases Fbh1 and Srs2 have nonredundant functionsin recombination (32, 33). An intriguing possibility is thatthe recombination role(s) performed by the sole Srs2 helicasein budding yeast, regulated by the PRR pathway and possiblyby phosphorylation (28), is achieved by distinct polypeptidesin fission yeast. Notably, S. cerevisiae Srs2, exclusively amongUvrD helicases, possesses a long C-terminal tail (Fig. 1B) containingmost of the DNA damage-inducible phosphosites (8), whose deletionalone is sufficient to suppress PRR mutants and which is requiredfor the physical interactions with both sumoylated PCNA (36)and Rad51 (26). If the functions of S. pombe Srs2 and Fbh1 couldbe reassumed by a sole PRR-regulated UvrD helicase in S. cerevisiae,then this would explain why the genetic relationships betweenSRS2 and PRR genes in S. pombe, or even in N. crassa, are lessconsistent with those depicted in S. cerevisiae. In fact, whilebudding yeast srs2 mutants suppress the DNA damage sensitivityof PRR mutants, S. pombe or N. crassa srs2 mutants do not (10,48). Furthermore, the same hypothesis might justify why thetiming of ubiquitylation- and sumoylation-dependent modificationsof PCNA, which are crucial for the regulation of Srs2 recombinationfunctions in S. cerevisiae (34, 36), has not been conservedequally in S. pombe (12). It is interesting that a phylogeneticanalysis performed on hemiascomycete yeasts revealed that theSRS2 locus is duplicated in tandem in certain species, and itis thought that gene duplication might facilitate their rapidevolution toward specialized functions (38).

In conclusion, our data are compatible with the idea that hFBH1is a bona fide human orthologue of S. cerevisiae Srs2 that,thanks to F-box-dependent functions, has evolved to executeits recombination role(s) independently of the PCNA modificationstatus, as instead happens in budding yeast.

It will be a challenge for the future to establish whether otherproteins functionally related to hFBH1 are present in humancells and whether inactivation of hFBH1 might have importantimplications for human health, like the case for the RecQ helicasesBLM, WRN, and RECQ4 (17).

ACKNOWLEDGMENTS

We are particularly grateful to M. Foiani for stimulating discussionsand for support. We thank S. Elledge, H. Klein, and H. Ulrichfor providing yeast strains and the Technological Services atIFOM for practical assistance. We also thank F. Asta, R. Bermejo,D. Branzei, W. Carotenuto, S. Confalonieri, J. Haber, H. Klein,C. Lucca, A. Pellicioli, T. Robert, R. Rothstein, and all themembers of our lab for helpful suggestions and/or comments onthe manuscript.

This work was supported by a grant from the Associazione Italianaper la Ricerca sul Cancro and by the European Community. I.C.is a recipient of an AIRC Unicredito Italiano fellowship.

FOOTNOTES

* Corresponding author. Mailing address: FIRC Institute of Molecular Oncology Foundation, Via Adamello 16, 20139 Milan, Italy. Phone: 39-02-574303306. Fax: 39-02-574303310. E-mail: giordano.liberi{at}ifom-ieo-campus.it

Published ahead of print on 27 August 2007.

Present address: Department of Genome and Computational Biology,Lawrence Berkeley National Laboratory, MS 84-171, 1 CyclotronRoad, Berkeley, CA 94720.

Present address: Banting and Best Department of Medical Researchand Department of Medical Genetics and Microbiology, Universityof Toronto, Toronto, Canada.

REFERENCES

Top